Sr. Engineer, Kernel Development and Optimization
Salary not listed • Belgrade, Serbia • Full-time
Posted 7mo ago
This job has been open for more than 60 days. It may no longer be actively hiring, so check with the company before putting a lot into an application.
Remote Work Policy
On-site
Employment Type
Full-time
Categories
Machine Learning Engineer
About the job
Tenstorrent is a leader in AI technology, developing next-generation AI compute solutions. The Kernel Development and Optimization team is crucial in unlocking the full potential of their hardware for ML and HPC workloads by creating performance-critical kernels. This role involves designing, implementing, and optimizing GPU-style kernels, contributing to host-side orchestration, and ensuring kernel correctness and performance through rigorous testing and collaboration with various engineering teams. You will gain deep insights into AI hardware, learn to optimize kernels outside traditional CUDA ecosystems, and explore AI-assisted workflows for kernel development.
Responsibilities
- Design, implement, and optimize GPU-style kernels (e.g., matrix multiplication, attention primitives, data-movement operations).
- Take ownership of performance, identifying bottlenecks and delivering measurable throughput improvements.
- Contribute to host-side orchestration code and parallelization strategies.
- Develop micro-benchmarks, regression tests, and tooling for correctness and sustained performance.
- Collaborate with compiler, runtime, ML, and hardware teams for kernel integration.
Requirements
- Strong C++ systems engineering background with experience in performance-critical or low-level software.
- Comfortable reasoning about concurrency, synchronization, latency hiding, and compute vs. memory trade-offs.
- Data-driven approach using profiling and benchmarking for optimization.
- Effective debugging of complex runtime or kernel-level issues.
- Structured thinking to break down performance problems into measurable experiments.
Benefits
- Highly competitive compensation package